I guess I should have given some background - I always hesitate to say too much (the water can lie still).
For months I've seen several Mandelbulber videos on YT, and thought "those are cool". Done 2D zooms with Tierazon for fun a while ago, and also other programs from many years ago. Something about a week ago or so "clicked" that I should download and run Mandelbulber - "just to see something". I can't remember exactly what it was. Maybe related to playing with TopMod recently too, and heavily working on some 3D mapping problem (that I still haven't worked out), for a video-game I'm working independently on. By "independently" I mean I'm not working for anyone, and stopped looking for work when the idea for this game came to me. It also means living on a shoestring budget while I wade my way through the work I have in front of me. The game is a puzzle very similar experience to a Rubik's cube - although completely unrelated.
After running Mandelbulber for a little while, of course some gorgeous images, and wondering if they could be incorporated somehow into the game. And very rapidly going from wondering... to knowing "these MUST go in". I hope everyone forgives me if I don't post many images - I might post some images later as some sort of promotion.
There are some hurdles I'm trying to overcome - which my first post points to:
1) How to rapidly/methodically find and reproduce nice locations in a formula? Right now I'm planning to write a Python script that creates keyframes that jump around overnight, and nice areas can be rendered later at higher resolution, etc. [Had written a script in sh, but hit a wall with floating point and glanced at Perl with much groaning and decided Python would be better to learn, if I have to learn anything.] For example, create 2000 keyframes, run the animation 1 frame/keyframe overnight at 32x32, then in the morning interesting images run those keyframes at higher resolution. [Or alternatively, open the timeline, which will take evernight to open with that many keyframes - and delete the ones that are not interesting directly in the timeline.]
Ideally keyframes could be generated in MB for any number of generated animation frames. Using the log output didn't work for me; the coordinates for every animation frame weren't enough to reproduce the images, and no good explanation. So the scripts generating keyframes is just a workaround.
2) Also rapidly/methodically finding formulas that are aesthetic but also not too expensive (higher DE is fast, but some formulas are unforgiving and require DE<0.1 - and even though they might be nice, I can't afford long render times). Could simply use another script that is running Mandelbulber repeatedly jumping over formulas.
3) I'm also very curious about using rented computing services (Linode, AWS) for rendering - if anyone is already using such resources please let me know how it's going.
Thanks for some of the great posts - tonight I was reading some and having a good laugh. weaver especially.
